查找具有最佳匹配评分的用户

Question

有没有人能帮我创建一种性能很好的方法来做以下事情？

我有一些用户评分一些不同的项目。

为了给用户提供好的推荐，我需要找到具有相似品味的用户。这样我就可以推荐其他有相似品味的用户的高评价商品。

我的数据来自与Entity Framework映射的数据库。

我有一个有效的算法，但性能很差，无法进行很多循环。有什么想法吗？

/kent

namespace ConsoleApplication1
{
    using System;
    using System.Collections.Generic;

    class Program
    {
        static void Main()
        {
            const int NoOfUsers = 100;
            const int NoOfRatingsPerUser = 100;
            const int NoOfDifferentItems = 100;

            var listOfRatings = new List<Rating>();

            var r = new Random();

            for (int user = 0; user < NoOfUsers; user++)
            {
                for (int ratingsPerUser = 0; ratingsPerUser < NoOfRatingsPerUser; ratingsPerUser++)
                {
                    var rating = new Rating();
                    rating.Score = r.Next(1, 10);
                    rating.ItemId = r.Next(1, NoOfDifferentItems);
                    rating.UserId = user;
                    listOfRatings.RemoveAll(x => x.UserId == user && x.ItemId == rating.ItemId);
                    listOfRatings.Add(rating);
                }
            }
            Console.ReadLine();
        }

        struct Rating
        {
            public int ItemId;

            public int Score;

            public int UserId;
        }
    }
}

Answer 1

我没看到评级算法。你在搜查的时候被杀了。KeyedCollection在这一点上非常适合。

在P4上为62毫秒。在1000 (而不是100)；

    public void BuildData()
    {
        System.Diagnostics.Stopwatch stopWatch = new System.Diagnostics.Stopwatch();
        stopWatch.Start();

        RatingKC Ratings = new RatingKC();

        const Int16 NoOfUsers = 1000;
        const Int16 NoOfRatingsPerUser = 1000;
        const Int16 NoOfDifferentItems = 1000;
        Int16 itemID;
        int score;
        Int32 key;
        var r = new Random();

        for (Int16 userID = 0; userID < NoOfUsers; userID++)
        {
            for (int ratingsPerUser = 0; ratingsPerUser < NoOfRatingsPerUser; ratingsPerUser++)
            {
                itemID = (Int16)r.Next(1, NoOfDifferentItems);
                score = r.Next(1, 10);
                key = (userID & 0xffff) + (itemID << 16);
                if (Ratings.Contains(key))
                {
                    Ratings[key].Score = score;
                }
                else
                {
                    Ratings.Add(new Rating(userID, itemID, score));   
                }
            }
        }
        stopWatch.Stop();
        // Get the elapsed time as a TimeSpan value.
        TimeSpan ts = stopWatch.Elapsed;

        // Format and display the TimeSpan value.
        string elapsedTime = String.Format("{0:00}:{1:00}:{2:00}.{3:00}",
            ts.Hours, ts.Minutes, ts.Seconds,
            ts.Milliseconds / 10);
        System.Diagnostics.Debug.WriteLine("Count = " + Ratings.Count.ToString() + " Time = " + elapsedTime);
    }
}
public class RatingKC : KeyedCollection<int, Rating>
{
    // The parameterless constructor of the base class creates a 
    // KeyedCollection with an internal dictionary. For this code 
    // example, no other constructors are exposed.
    //
    public RatingKC() : base() { }

    // This is the only method that absolutely must be overridden,
    // because without it the KeyedCollection cannot extract the
    // keys from the items. The input parameter type is the 
    // second generic type argument, in this case OrderItem, and 
    // the return value type is the first generic type argument,
    // in this case int.
    //
    protected override int GetKeyForItem(Rating item)
    {
        // In this example, the key is the part number.
        return item.Key;
    }
}
public class Rating: Object
{
    public Int16 UserId { get; private set; }
    public Int16 ItemId { get; private set;}
    public int Score { get; set;}
    public Int32 Key { get { return (UserId & 0xffff) + (ItemId << 16); } }
    public override bool Equals(Object obj)
    {
        //Check for null and compare run-time types.
        if (obj == null || !(obj is Rating)) return false;
        Rating item = (Rating)obj;
        return (UserId == item.UserId && ItemId == item.ItemId);
    }
    public override int GetHashCode()
    {
        return Key;
    }

    public Rating(Int16 userId, Int16 itemId, int score)
    { UserId = userId; ItemId = itemId; Score = score; }
    public Rating(Int16 userId, Int16 itemId)
    { UserId = userId; ItemId = itemId; }
}